Evaluation of effects of small-incision approach treatment on proximal tibia fracture by deep learning algorithm-based magnetic resonance imaging

Abstract

Abstract In this study, magnetic resonance imaging (MRI) based on a deep learning algorithm was used to evaluate the clinical effect of the small-incision approach in treating proximal tibial fractures. Super-resolution reconstruction (SRR) algorithm was used to reconstruct MRI images for analysis and comparison. The research objects were 40 patients with proximal tibial fractures. According to the random number method, patients were divided into a small-incision approach group (22 cases) and an ordinary approach group (18 cases). The peak signal-to-noise ratio (PSNR) and the structural similarity index (SSIM) of the MRI images before and after the reconstruction of the two groups were analyzed. The operative time, intraoperative blood loss, complete weight-bearing time, complete healing time, knee range of motion, and knee function of the two treatments were compared. The results showed that after SRR, the PSNR and SSIM of MRI images were 35.28 and 0.826 dB, respectively, so the MRI image display effect was better. The operation time in the small-incision approach group was 84.93 min, which was significantly shorter than that in the common approach group, and the intraoperative blood loss was 219.95 mL, which was significantly shorter than that in the common approach group (P < 0.05). The complete weight-bearing time and complete healing time in the small-incision approach group were 14.75 and 16.79 weeks, respectively, which were significantly shorter than those in the ordinary approach group (P < 0.05). The half-year knee range of motion and 1-year knee range of motion in the small-incision approach group were 118.27° and 128.72°, respectively, which were significantly higher than those in the conventional approach group (P < 0.05). After 6 months of treatment, the rate of good treatment was 86.36% in the small-incision approach group and 77.78% in the ordinary approach group. After 1 year of treatment, the rate of excellent and good treatment was 90.91% in the small-incision approach group and 83.33% in the ordinary approach group. The rate of good treatment for half a year and 1 year in the small incision group was significantly higher than that in the common approach group (P < 0.05). In conclusion, MRI image based on deep learning algorithm has a high resolution, good display effect, and high application value. The small-incision approach can be applied to the treatment of proximal tibial fractures, which showed good therapeutic effects and a high positive clinical application value.